Construction workers not only have higher rates of work-related injuries than do other trade groups, but they are also among the most likely workers to experience serious occupational injuries. Their fatal and lost-work-time injuries continue to rank among the highest in the United States, and national rates of disabling injuries have risen in the rank among the highest in the United States, and national rates of disabling injuries have risen in the construction trades in recent decades. Construction work is transient in nature, often involving small and dispersed work sites, and these workers frequently change employers. All of these factors necessitate the development of creative ways to study these historically difficult-to-reach high-risk workers. The primary objectives of this study are first to develop methods for cataloguing and coding data appearing in injury incident descriptions found in standard First Reports of Injury (FRI) and Accident Investigation Reports (AIR) including free text, and to create a database to link these data to existing coded administrative data. The combination of information from injury reports with an administrative database containing claims, demographic information and hours worked will provide a rich source of information for describing injuries and the factors contributing to them. The utility of these data will then be evaluated to test specific hypotheses about factors differentially associated with different types of injury as well as different levels of injury severity. The data will also be used to determine direct costs of injury for high risk groups of workers and for specific factors contributing to injury. Together, these outputs are designed to identify causes of work- related injury and to focus prevention efforts on more costly risk factors. Study aims will be accomplished by coding and analyzing data from standardized FRIs and AIRs, linked with an administrative database for the Denver International Airport construction project containing information on over 4,600 workers' compensation claims; payroll and demographic data on 32,081 workers who worked over 31 million hours on the project; and company characteristics for 769 contractors. The results should provide information that will be useful in focusing specific prevention efforts for construction, but the methods developed will have applicability to data for other occupational groups as well.